Catalan researchers create "new switches" to control anything with AI
The tool created by researchers at the Center for Genomic Regulation allows the creation of synthetic DNA fragments with customized criteria.
"We have taken a very important step towards understanding the language of DNA." With this statement, Lars Velten, a researcher at the Center for Genomic Regulation (CRG), tried to quantify the extent to which the results of the study his team published this Thursday in the journal CellThe research group from the CRG and EMBL-Barcelona has succeeded in creating an artificial intelligence (AI) tool capable of designing regulatory DNA sequences never seen in nature. "We can say that we have created new switches [enhancers] that until now did not exist in the mammalian genome and that in the future could be used to control whether a gene is turned on or not in a specific cell," explains Velten, lead author of the research.
The tool created by the CRG team allows, using an AI model, to create synthetic DNA fragments with personalized criteria. In this way, for example, the model could be asked to give the new DNA fragment the order to activate a specific gene in stem cells so that they become red blood cells but not platelets. "The potential applications are enormous. It's like writing software, but for biology. It gives us new ways to give instructions to a cell and guide the way it develops and behaves with unprecedented precision," explains Robert Frömel, first author of the study.
Broadly speaking, this new AI model predicts which combination of DNA letters (A, T, C, G) is needed to activate certain genes in specific cell types. This information helps researchers chemically synthesize DNA fragments of approximately 250 letters and add them to a virus for delivery into cells.
To test the new tool, the researchers asked the AI to design synthetic fragments that would activate a gene encoding a fluorescent protein in certain cells. The fragments created from scratch were inserted into mouse blood cells, where the synthetic DNA was fused with the genome at random locations, and the experiment worked as planned.
Know what causes certain diseases
The CRG study marks a milestone in the field of generative biology, as it could help develop new gene therapies that increase or decrease gene activity where needed. It also paves the way for new strategies to fine-tune a patient's genes and make treatments more effective and reduce side effects.
Furthermore, Velten explains that, in the future, tools like his could be key when the cause of a disease in a patient is unknown. According to the researcher, the tool could "interpret genome sequences to identify the variants and mutations" that can cause it. However, Velten acknowledges that this is "a long-term goal," since "the genome is very complex" and has thousands of proteins that can interact with the enhancers that activate genes, and so far, they have focused on only 40 of these "switches."